Oskar A. Haas
Cancer can be seen as the consequence of a chaotic process, a combination of Murphy's Law and Darwin's Law* anything that can go wrong will, and in a competitive environment, the best adapted survive and prosper.
--W. Wayt Gibbs, Scientific American July 2003, p. 49
Setting the stage
The healthy human body is a complex linked system of cells that is composed of several dynamic equilibria whose disturbances result in disease. In this context, cancer can be viewed as a genetic disease deriving from the territorial expansion of a mutant somatic cell clone, which disrupts the organism's physiologic balance between ordered cell production, differentiation, and elimination. Consequently, the characteristic pathology of the neoplasm results from the interaction between the expanding mass and its specific somatic environment.1 Thus mutations that collectively corrupt cellular control pathways of a single immature or undifferentiated precursor cell can be viewed as the ultimate cause of cancer. However, cancer is not a deterministic genetic disease because continuous destabilization of the genome during the selfperpetuating process of neoplastic transformation also involves characteristics, physiologic factors, and exposures that not only induce but also influence the probability of mutation. Thus the ensuing concept of a multistep evolution of cancer implies that spurts of mutations propagate tumour development via a Darwinian process of natural selection whose essential components are cell proliferation, genetic diversification, environmental pressure, competition, and adaptation.2 The important point here, which is commonly overlooked, is that the internal and external environments influence and modify the behaviour of genes as much as genes themselves determine their intracellular and extracellular milieu. Therefore the environment plays a significant role in the selection of mutations that are appropriately adapted for the successful survival of genetically destabilized cells.3
The internal and external homeostatic conditions of a developing host differ significantly from those of a fully mature one. Therefore it is not surprising that the spectrum of haematologic malignancies and solid tumours, as well as their biologic features and response to therapy, varies markedly between children and adults. The vast majority of childhood malignancies result from errors that take place during early stages of cell differentiation, tissue maturation, and organ development.4,5 The cellular self-organization into organs during embryonic development is driven by cell–cell communication. Disturbances in this process often result in disease. Genetic destabilization of productive progenitor cells creates a large amount of heterogeneity. Particular circumstances can provide one or the other of the affected cells with increased fitness and survival advantages, which are the first steps on the path to tumour development. Additional factors that contribute to the formation of the diversity of tumours in children and adults include a different susceptibility in the stem cell population that is injured by mutations as well as the type and numbers of mutations that are necessary to induce a fully malignant phenotype. Thus the predominance of specific types of mesenchymal tumours may reflect a combination of the high cell turnover during early and mid-fetal development that makes certain precursor cell populations more vulnerable to DNA replication errors, and the limited number of mutations that are necessary for the induction of overt disease. Moreover, the number and nature of cooperating mutations required to induce a fully malignant phenotype probably also varies depending on the initiating lesions. The low frequency of adult types of epithelial neoplasms in children, on the other hand, is in agreement with the idea that their development requires several more genetic alterations and steps and consequently a much longer time to become clinically apparent.5,6
Rather than listing the specific cytogenetic and molecular changes that characterize individual types of childhood malignancies in this chapter, we intend to focus more on these conceptual aspects of the role and significance of genetics in childhood cancer. Several unusual views and inspiring theories that have been put forward recently provide the basis for exploring how, when, and where mutations are generated, how particular patterns of genetic abnormalities emerge, in which way they reflect particular internal and external environmental conditions, and why recurrent patterns of genomic imbalances are so specifically linked with distinct types of tumours in both children and adults. In particular, concentrating on childhood malignancies illustrates that much can be learnt from studying rare conditions.7The relevance of the clues stretches far beyond the particular disease entity. Thorough analyses of genetic alterations, especially those occurring in haematologic neoplasms in infants and twins, have recently provided some striking insights into the time course and the potential steps and mechanism of the transformation process.7,8 This knowledge is essential for understanding the biology, pathology, and clinical behaviour of the ensuing tumours and eventually for devising more effective and less harmful treatment strategies. Moreover, as an added bonus, genetic studies of childhood malignancies also help to elucidate the normal physiologic developmental processes of cells, organs, and the human organism. Last but not least, they also add to understanding of the pathogenesis of adult tumour forms.
Genetic information
The information provided in this section derives mainly from the very instructive and inspiring article by Hood and Galas,9 as well as from a compilation of several other recently published excellent review articles on this topic.10,11
Processing and utilization
Cells, organs, and organisms consist of heterogeneous components that are organized in distinct levels and interact within large networks. At the core of biologic complexity resides DNA, the immortal molecule of life that encodes genomic information and fuels the process of molecular self-organization. Although the DNA structure provides an immediate explanation for mutation and variation, change, species diversity, evolution, and inheritance, it does not automatically provide a mechanism for understanding how the environment interacts at the genetic level.11
The two main elements of genomic information operate across three time spans: evolution, development, and physiology. Genes encode the RNA and protein machinery of life, whereas the regulatory networks define how these genes are expressed in time, space, and amplitude. The regulatory networks consist of two major components: the transcription factors, and the sites in thecontrolregionsofgenestowhichtheybind, suchaspromoters, enhancers, andsilencers. Longterm information is stored almost exclusively in the genome and short-term information in the proteome. As becomes evident from the changing morphologic patterns during the development from a fertilized egg to an adult organism, it is not so much the genes, but rather the regulatory networksthat play the crucial role in this process. Consistent with this view, individual cellsutilize only the very small part of the genome that is necessary to maintain their function, development, and differentiation at a given time point, and by far the largest proportion of their genome always remains inactive. This also applies to neoplastic tissues, although it must be remembered that in this situation the genome is additionally continuously perturbed and rearranged.
The DNA-encoded information flows from a gene to the environment in a distinct fashion which is subject to a variety of regulatory feedback loops and environmental influences. At each successive layer, the information becomes more complex and information can be added or altered for any given element. This includes, for example, the flexible remodelling of the chromatin configuration, RNA splicing, and protein modification (Fig. 2.1). The further the flow of information drifts away from the DNA, the more it is influenced and modified by the intra-and extracellular environment, and especially by other cells or chemical gradients. The elementary building blocks organize themselves into recurrent patterns and functional modules that shape the discrete cellular functions, such as gene regulatory circuits and metabolic pathways. The cellular interactions are cooperative and self-enhancing, and their autocatalytic feedback properties provide the means of altering strategies and adapting to changes in the internal and external environments. Moreover, cells and organisms also modulate the strength of the interaction with neighbouring units to increase their fitness, the maximum of which occurs at the boundary of order and chaos.
|
|
|
Fig. 2.1 The world according to DNA. The DNA of an individual cell encodes all the information that is necessary to form the complexity of the whole organism. Conversion of information takes place at several distinct levels and is a plastic and flexible process that is subject to many feedback loops and both intra- and extracellular environmental influences. |
During the embryonic and fetal periods, cells increase dramatically in number, mature, and become specialized to form tissues and organs. An adult human body consists of an estimated 1014 cells and approximately 300 distinguishable cell types. During a human lifetime approximately 1017 cells are produced in total. Two hundred million cells are lost and renewed per minute, which is equivalent to a cell turnover of one body mass every 35 days. Thus this highly ordered and structured agglomeration of cells exists in an intricately labile balance. This permanent consumption and loss of cells has to be compensated by continuous cell renewal processes, whose starting points are highly specialized tissue-specific stem cells. The derived daughter cells differentiate along well-defined pathways. To counter cell production and maintain the necessary steady state, the controlled physiologic elimination of cells by apoptosis occurs in both the fetus and the adult. Intriguingly, however, and as will be explained later, apoptosis also has the catastrophic potential to promote mutation and genome instability. Therefore the fundamental idea is that a combination of deregulated proliferation and suppressed apoptosis constitutes the minimal platform upon which all neoplasms reside.1,2
Epigenetic inheritance and imprinting
Genetically determined differences in gene activity are usually attributed to DNA sequence variations, such as polymorphisms and mutations. However, the facts (e.g. the phenotypic traits of twins and cloned animals, with their supposedly identical DNA make-up, can still vary considerably) imply that a DNA-sequence-independent and environmental modulation of the transcribed genetic information must be at least as important as the original DNA-derived message itself. The main player in this process is the dynamic plasticity of the chromatin organization, which exerts a profound control over gene expression and other fundamental cellular processes. Such heritable, but DNA-sequence-independent, regulatory mechanisms of gene expression govern the flexible conversion of the DNA-encoded information into cell- and tissue-specific gene activity patterns. This is summarized under the umbrella term ‘epigenetics’.12
The most important regulator is the reversible methylation of gene promoters and the inextricably linked methylation and acetylation of the associated histone proteins. Although these modifications can be altered by a variety of factors, they can also remain remarkably stable and be passed on to future generations. This kind of cellular memory ensures that fundamental decisions regarding turning individual genes or groups of genes on or off need to be made only once and that daughter cells inherit identical repressed or activated transcription states. This provides an efficient mechanism for cellular differentiation. Since such patterns of epigenetic information are also replicated, mutated, and selected in the somatic environment, they should, like the DNA sequence itself, also evolve by Darwinian mechanisms.3
The classic example of the phenomenon of epigenetic inheritance at the generational level is imprinting. This type of non-Mendelian inheritance of specific traits results from the parent of origin-specific transmittance and the parent of origin-specific monoallelic usage of specific genes. Since many of these genes encode fetal growth factors and their receptors, imprinting plays an important role in fetal development. Paternally expressed genes enhance fetal growth, whereas maternally expressed genes suppress it. In mammals, for example, the gene encoding insulin-like growth factor 2 (IGF-2) is only expressed from the paternal copy of the gene, whereas the H19 gene is expressed solely from the maternal allele. The uncoordinated expression of these and other genes can cause various developmental disturbances and diseases, including overgrowth and tumour-predisposition syndromes, the prototype of which is the well-known Beckwith–Wiedemann syndrome.13
However, epigenetic mechanisms play a much more profound role in neoplasia and are instrumental in virtually every step in its initiation and progression. Patterns of DNA methylation and chromatin structure are severely altered and include genome-wide hypomethylation as well as regional hypermethylation of specific promoters.12
DNA integrity, DNA repair, and cell-cycle checkpoint control
The preservation of its genomic information is of paramount importance for a cell. This task requires sophisticated surveillance and maintenance machinery, because DNA is a highly thermolabile molecule that is permanently jeopardized by a variety of internal and external noxious agents.14 The two essential parts of this machinery are the DNA damage checkpoint system and the DNA repair enzymes.2,11 The complex surveillance system somehow senses genome injury and arrests the cell cycle at specific checkpoints in the G1,S,G2, and M phases, activates DNA repair networks, or induces programmed cell death (apoptosis).
The two major categories of mechanistically distinct events of genome damage are those which affect chromosome numbers, and those which alter chromosome structure. The former most likely reflect malfunction of the mitotic chromosome segregation apparatus, whereas the latter point to irregularities in DNA repair processes. Potential contributors to chromosome mis-segregation include centrosome dysfunction, anaphase checkpoint malfunction, and cytokinesis failure.15
Activation of the repair system not only stimulates transcriptional programmes and the movement of the DNA repair proteins to sites of DNA damage, but also regulates the telomere length. Thus the potential sources of genomic destabilization comprise defective copying, excessive damage, ineffectual repair, and faulty segregation. The outcome of DNA damage is diverse and generally harmful. Severe damage is lethal to the cell, or at least severely deleterious to its proliferation. Insufficient repair, on the other hand, converts less severe DNA corruptions into permanent mutations, whose irreversible long-term effects contribute to ontogenesis.
In an analogous fashion, malfunction of the telomeres also adds significantly to chromosome breakage, fusion, and mis-segregation. Telomeres consist of arrays of repetitive sequences that protect the chromosome ends from degradation during cell growth and differentiation. Gradual shortening during successive cell divisions eventually results in crisis and cell death. Neoplasms circumvent this fate both by inactivating the cell death pathway and by switching on telomerase. This enzyme helps to maintain telomere length, thereby rendering the cell immortal.16,17 Thus passage of cells through a crisis in the setting of deactivated DNA damage checkpoints provides a mutational mechanism that can generate a variety of cancer-initiating genetic alterations.17 Although the replicative potential of tumour-forming precursor cells in infants and children is still considerably higher than that of the epithelial cells of aging adults, telomerase is nevertheless activated in a variety of childhood tumours. One explanation for this phenomenon might be that in this setting it may be necessary to maintain the proliferative potential of a tissue that has already been transformed rather than trigger tumour initiation by rescuing senescent cells, as is the case in adults. This would also explain the much more pronounced karyotype instability and complexity of chromosome rearrangements in the epithelial tumours of adults compared with those in the mesenchymal tumours and leukaemias of the young.
Allocation of an appropriate repair system is not a trivial process, as the choice depends on where and in which physiologic context DNA strand breaks occur in the genome.2,11,14 Therefore a particular DNA repair pathway may be appropriate or normal in one situation, but inappropriate or abnormal in another. Point mutations arise from faulty repair of DNA base damage, mis-incorporated DNA bases, or spontaneous or induced deamination of methylated cytosines. Small-scale DNA damage is reversed through base excision, nucleotide excision, and mismatch repair, whereas DNA double-strand breaks (DSBs) can be repaired with high fidelity through homologous repair. Unrepaired and unprotected DSBs, including exposed telomere ends, are fusogenic and joined to other double-strand DNA ends. Inappropriate repair by the non-homologous end-joining (NHEJ) system generates structural gene and chromosome rearrangements. Repetitive interor intragenic simultaneous or consecutive breakage-fusion–breakage cycles may then lead to very complex structural rearrangements. The fact that break-originating processes, such as deletions, insertions, inversions, amplifications, and translocations, are particularly abundant in cancer indicates that this pathway of NHEJ is frequently involved. It acts as an emergency repair of broken chromosomes, but at the cost of fidelity. This is further corroborated by the increased incidence of tumours in humans with repair deficiency syndromes and is also backed up by experimental evidence from knockout mice lacking the corresponding genes.15
In addition, irregularities that occur during DNA synthesis and replication can also generate substrates for recombination and give rise to gross chromosomal aberrations. Thus, in the absence of external sources of DNA damage, the increased formation of destabilizing DSBs can also be triggered by endogeneous metabolic defects or enhanced irregular DNA excision repair. The preliminary observations that exposure of preterm infants to diagnostic X-rays did not increase leukaemogenic fusion transcripts whereas epidemiologic evidence points to an association between neonatal oxygen supplementation and leukaemia are of particular interest in this context. Thus gratuitous repair may be an important source of spontaneous mutation.
Finally, failure in DNA damage signalling is probably at least as harmful as a specific DNA repair defect. It allows mutagenic repair of DNA damage while lowering the rate of apoptotic cell death. Thus, in terms of the costs and benefits of DNA repair, stopping for repairs can be a fatal strategy in a hostile environment.3 For instance, concurrent defects in the nucleotide excision repair and cell-cycle control systems would provide such a situation in which repairs could take place without stopping. It is a destabilizing and risky venture, but it might be the only option in a mutagenic ‘war zone’, as expressed so vividly by Breivik in a recent article.3
Genetic susceptibility and predisposition
Although the vast majority of childhood malignancies appear to occur sporadically, their early age of onset suggests that underlying inherited constitutional genetic defects might provide a strong predisposing influence.18,19 Overall, more than 600 genetic predisposition factors have been recognized so far. They include rare high-penetrant DNA mutations and epigenetic defects as well as more common genetic polymorphisms that influence individual response to environmental exposure.20 The classic constitutional mutations in children include sporadic chromosome anomalies as well as chromosome and gene defects that are inherited according to the Mendelian laws, i.e. autosomal dominant, autosomal recessive, and X-linked traits. More common genetic traits, such as those that influence the metabolic activation or detoxification of carcinogenic chemicals, are probably important determinants of population risk although they pose low individual risk. Intriguing mouse breeding experiments have led to the provocative speculation that offspring of parents with relatively divergent DNA sequences might be less susceptible to mitotic recombination and therefore more protected from cancer than those of less divergent parents.
Furthermore, the age-related differences that exist in susceptibility to environmental toxicants are noteworthy. Experimental and epidemiologic data indicate that, apart from predisposing genetic traits or ethnicity, the very young may have a heightened risk because of their differential exposure and physiologic immaturity.18 For example, relative to body weight, infants and children take in appreciably more food, water, air, and any carcinogens contained therein than adults. Moreover, increased absorption and retention of toxicants, reduced detoxification and repair, and the higher rate of cell proliferation may constitute additional factors that contribute to the higher internal dose of toxicants and greater genetic damage to infants and children compared with similarly exposed adults.18 Moreover, preliminary observations also suggest that the risk of leukaemia initiation in utero may be modified by dietary as well as genetic factors.7
Types and consequences of cytogenetic and molecular genetic alterations
Gene, chromosome, and genome mutations
Human neoplasms exhibit genome modifications that range from subtle point mutations to dramatic gains and losses of whole or parts of chromosomes.21 Such chromosome aberrations are the visible hallmarks of gene deregulation and genome instability.2 It is noteworthy that acquired cancer-associated mutations never occur constitutionally. Three distinct types of genome instability contribute directly or indirectly to the unbalanced expression of genes: chromosome instability, microsatellite instability, and epigenetic instability (discussed above).
The term ‘chromosome instability’ includes not only gains and losses of entire chromosomes, but also gains and losses of chromosome parts (deletions and duplications) and amplifications as well as structural rearrangements (translocations, inversions, and insertions).15 Microsatellite instability, on the other hand, results from a defective mismatch repair system, which particularly affects DNA repeat sequences and causes surprisingly little chromosome instability. It is a specific hallmark of various inherited and sporadic forms of adult cancers, but is not instrumental in the pathogenesis of childhood forms of cancer.
One of the hotly debated questions at present is whether chromosomal aneuploidy or gene mutation comes first, and which type of aberration matters most.15 A compromise suggestion is that aneuploidy may collaborate with intragenic mutations during tumorigenesis by altering the dosage of thousands of genes, thereby accelerating the accumulation of oncogenes and the loss of tumour suppressor genes. In contrast with the most common forms of carcinomas in the adult, the genome of leukaemia, lymphoma, and sarcoma is not as scrambled and often remains quite stable and within the diploid range. This distinction is important, because it implies that such specific changes result from an early single hit, whereas the more complex and clonally unstable changes reflect a more pronounced underlying defect in chromosome and genome maintenance. Even aneuploidy exhibits relatively stable chromosome complements in these malignancies, probably because it arises from a rare event in a tumour founder cell rather than from a catastrophic destabilization of the genome in senescent cells, as is generally the case in carcinomas.15
Aneuploidy in childhood malignancies
Three extraordinary types of aneuploidy are of particular interest: constitutional trisomy 21, hyperdiploid acute lymphoblastic leukaemia (ALL), and pseudotriploid neuroblastoma.
Constitutional trisomy 21 is a condition that predisposes to a self-limiting transient myeloid proliferative disease (TMD) as well as to progressive haematologic malignancies.22 TMD occurs exclusively during the postnatal period, whereas true leukaemias develop later, within the first few years of life. The most common form is acute megakaryocytic leukaemia (AML-M7). TMD and AML-M7 are related clonal diseases which, except for the age of onset, are virtually indistinguishable by any currently available diagnostic measures. Although the responsible factor was suspected to reside on chromosome 21 itself, it was recently proved that mutations of the X-linked GATA1 gene are the culprit. Since these mutations occur in the fetal liver and arrest the differentiation of the megakaryopoietic lineage, they are found in both TMD and AML-M7 and are the long-sought cause of leukaemic transformation.22
Although the mechanism leading to the increased chromosome number in hyperdiploid ALL with 51–65 chromosomes still remains unknown, there is good evidence that a single nondisjunction event must lead to the non-random gain of the trisomic chromosomes 4, 6, 10, 14, 17, 18, 20, and X and the tetrasomic chromosome 21.23 Once formed, the abnormal karyotype remains remarkably stable. The ensuing imbalances may either enhance the proliferation capacity of early lymphoid cells solely through a change in dosage or relative dosage of a set of genes or, in a similar process, block differentiation. This has recently been supported by gene expression studies, which have revealed that genes on chromosomes X and 21 are comparatively more expressed than others. The reason that specific gene mutations are not considered very important in this type of disease also derives from the fact that they generally lack structural rearrangements.24
The extraordinary biology and behaviour of neuroblastoma ranges from life-threatening progression to maturation into ganglioneuroblastoma and spontaneous regression.25 The latter is one of the most unusual aspects of infants with stage IV-S, a disseminated disease form that may involve liver, skin, and/or bone marrow, but not cortical bones or distal lymph nodes. Genetically, these tumours are characterized by a pseudotriploid karyotype that usually consists of pure non-random numerical changes. Specifically, they also lack the typical markers for progressive and late stages, such as MYCNamplification and 1p deletions. Moreover, such favourable neuroblastomas rarely, if ever, evolve into unfavourable ones. When, how, and why pseudotriploidy arises remains as enigmatic as its role in the unusual disease process.
Cancer genes and deregulated signal transduction pathways
Cancer-associated abnormalities disturb the regulatory circuits that govern normal cell physiology and homeostasis.2 The dominant paradigm is that four to ten mutations need to disrupt the right genes to alter five or six different regulatory systems on the path to malignancy (Table 2.1).6 The most important gene classes that are implicated in cancer are the gatekeeper genes that control cell growth and death and the caretaker genes that maintain genome integrity.
Members of either class can act as oncogenes or tumour suppressor genes. Cellular protooncogenes normally function as positive regulators of cell growth, whereas tumour suppressor genes serve as negative ones. Members of both classes interact and cooperate in distinct signal transduction pathways, one or the other of which is disrupted in virtually every benign and malignant human neoplasm.2 The most prominent examples of characteristic and specific karyotype and gene rearrangements encountered in childhood malignancies are listed in Table 2.2.
|
Table 2.1. Six essential alterations in cell physiology that collectively dictate malignant growth6 |
||||||
|
Oncogenes are frequently activated by gain of function mutations or fusions with other genes. They may also become aberrantly activated through amplification, increased promoter activity, or protein stabilization. The best-known examples of gain of function mutation are probably those of the RAS genes. Reciprocal chromosome rearrangements result in the illegitimate recombination or juxtaposition of two normally separate genes.8 The ensuing fusion genes generate either a hybrid mRNA and a chimeric protein with novel properties, such as an altered transcriptional regulation or activated kinase, or deregulate a partner gene. The former type of fusion is common in immature forms of B-cell precursor ALL, AML, and sarcomas, whereas the latter is a typical feature of mature B- and T-cell acute leukaemias and lymphomas. In the case of the formation of a chimeric fusion gene, this first or initiating event needs to be complemented with mutant or activated kinases, such as FLT3 tandem duplications or c-KITmutations, to lead to malignant transformation.2,11 The cytogenetic manifestations of high-level amplifications are double-minute chromatin bodies (DM) or homogeneously staining regions (HSR). Such gene amplifications cause an overexpression of the affected gene(s). The prototypic example from childhood malignancies is the MYCN amplification in neuroblastoma, which is a hallmark of advanced stages with a poor outcome.25
Tumour suppressor genes, on the other hand, are inactivated by physical loss (deletion), loss of function mutations, or epigenetically by promoter methylation. Whereas a mutation in one allele is enough to activate an oncogene permanently, both alleles usually have to be knocked out to inactivate a tumour suppressor gene completely.26 Knudson's classic two-hit model26 derived from his studies of hereditary and acquired forms of retinoblastoma and stipulates that the inherited or acquired mutation or loss of a single allele of a tumour suppressor gene should lead to a reduction of the gene dosage. In this model, the diminished expression remains below the tissue-specific thresholds that could interfere with the control of fundamental cellular processes. According to the model, tumours develop only when the function of the second allele is also lost. However, in other instances a monoallelic disruption of such a tumour suppressor gene alone may be sufficient to create a cellular phenotype that promotes tumour genesis, without the necessity of an inactivation of the second allele. Such a haplo-insufficiency of particular caretaker genes may immediately result in defective DNA repair and increased genetic instability and thereby induce somatic mutations in other tumour suppressor genes and oncogenes in a much shorter period of time. However, tumours influenced by haploinsufficiency usually have a later age of onset when compared with those caused by biallelic inactivation.27
|
Table 2.2. Representative examples of the most relevant chromosomes and gene rearrangements encountered in childhood haematologic neoplasms and solid tumours |
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Novel insights into the origin of reciprocal gene rearrangements
Chromosome abnormalities initiating childhood ALL might arise spontaneously and frequently as accidental by-products of the endogenous proliferative, apoptotic, or metabolic stress of haemopoiesis.8 The production of a functional chimeric fusion gene requires not only the simultaneous presence of DSBs in two chromosomes, but also their proximity at a particular time point, perhaps in a repair complex.8Although DSBs can occur randomly throughout the genome, they may nevertheless cluster in particularly vulnerable regions, such as open chromatin configurations and chromosomal scaffold attachment sites. The increased incidence of secondary leukaemias with an MLL gene rearrangement in children who had previously been treated with topoisomerase II inhibitors instigated experiments that helped to resolve this issue. They revealed that DNA cleavage at this particular locus results from chromatin fragmentation during the initial stages of drug-induced apoptosis. These findings provided the basis for the fascinating hypothesis that such reciprocal rearrangements are commonly generated by rescuing early stage apoptotic cells by means of emergency NHEJ repair.8 It also helped to partly explain the non-randomness of such recurrent rearrangements and the fact that they can be generated by very heterogeneous stimuli, such as increased cell turnover during physiologic B-cell development, cytostatic drugs, radiation or even viral infections. The detection of DNA-sequence patterns that are typical for NHEJ repair at the respective fusion sites render it likely that this is the predominant route for the generation of the majority of leukaemia- and sarcoma-associated chimeric fusion genes.
Translocations that involve the immunoglobulin (IG) and T-cell-receptor (TCR) loci, on the other hand, lead solely to a constitutive overexpression of the respective fusion partner without altering its gene structure; for example MYCC in the case of the Burkitt-lymphoma-associated t(8;14)(q24;q32). In contrast with the chimeric fusion genes, these types of gene fusions are commonly mediated by the RAG proteins and V(D)J recombinase. Under physiologic conditions, these enzymes promote somatic recombination in precursor B and T cells, which is a prerequisite for the generation of the IG and TCR repertoire.
More than 600 reciprocal chromosome rearrangements have been identified so far. They are highly specific markers for particular subentities of haematologic malignancies and solid tumours. The breakpoints of more than 285 such translocations and inversions have already been cloned and the approximately 280 genes involved have been characterized. Interestingly, only a small number of genes are frequently involved, whereas a large number are involved in a few or single rearrangements (Table 2.3).
Analysis and evaluation of genetic alterations
Particular features of neoplasms can be studied at any of the levels outlined in Figure 2.1.2 In particular, the genome itself and several of the subsequent levels can be investigated with two different, but complementary, approaches, namely morphologically using microscopy and chemically using molecular genetic techniques.16,21 The morphologic methods include all types and variations of light and fluorescence microscopy techniques, such as conventional cytogenetics and fluorescence in situ hybridization (FISH). On the other hand, all molecular genetic methods aim to obtain information about the composition of a particular DNA or RNA sequence for comparison with a known reference.9 Although the large number of methods are based on a few general principles, most of them have evolved into highly complex technologies whose application usually requires very sophisticated and expensive technical equipment. Examples of these methods are Southern blot analyses, as well as all types of polymerase chain reactions (PCRs) and DNA sequencing. Separation of the DNA- and RNA- based FISH methods from the cellular and nuclear topologic context led to the development of microarray or chip technology, which simultaneously allows the comparative analysis and evaluation of several thousand sequences.10,28 The expectation is that technologies such as bacterial artificial chromosome (BAC) microarrays for the high-resolution detection of genetic changes and cDNA-based, oligonucleotide-based, or high-throughput proteomics approaches for the detection of gene expression will eventually facilitate the characterization of the complex network of interactions that are associated with the development of neoplasms.2,16 However, it has been pointed out that these spectacular technologies are seductive and sometimes corrupting, and one should not commit the mortal sin of genomics and confuse throughput with output and data with knowledge.10
|
Table 2.3. The top 10 most frequently involved genes |
||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||
The selective forces of the internal and external environment
The two sides of mutagens and carcinogens
The external environment always exerts its influence on the DNA level in combination with genetic and acquired susceptibility.11,18 Cancer-promoting agents not only leave their footprints on DNA, but also shape the somatic evolution of the entire genome of cancer cells.3 Epidemiologic surveys together with clinical observations and intricate genetic analyses provide an increasing variety of illustrative examples for this idea, particularly in leukaemias of infants and young children.8 Although it is becoming increasingly evident that genetic instability is a non-random event, the currently prevailing assumption is that mutagens and carcinogens damage the genome randomly and that the environmental risk for cancer is a direct consequence of exposure to environmental mutagens. However, this proposal has recently been challenged.3,29 Based particularly on analyses and comparisons of in vitro and in vivo incidence and distribution rates of point mutations in various tumours and cell cultures, several researchers now consider that this originally plausible hypothesis is, to a large extent, unsupported by evidence. They believe that the contribution of environmental factors to point mutagenesis is negligible, except for very specific circumstances, such as exposure to solar radiation, chemo- and radiotherapy, or the in vitro exposure of cells. As an alternative, they suggest that most tumour-associated mutations result primarily from endogenous processes, such as errors during the turnover of undamaged DNA, whereas environmental conditions would be more likely to select than induce such ‘oncomutations’.3,29 One of the arguments is that the likelihood of DNA polymerase errors remains the same, irrespective of whether DNA bases are excised from damaged or undamaged DNA. Moreover, a DNA turnover rate of 1% per cell per day would require much more intensive DNA repair than the excision repair of even a million DNA damage sites.
At first glance this novel view of the role of mutagens appears counter-intuitive, but on further consideration it seems quite plausible. It fits very well with what we have discovered within the last few years about the molecular foundation of fusion genes. Regardless of the specific mechanism of their induction—be it chromosome breakage, cellular endonucleases, chemical interactions with topoisomerase II inhibitors or RAG-mediated cleavage-the final common denominator is an endogenous process, the NHEJ repair.8
The significance of (cancer) stem cells
Although tumour-initiating mutations may theoretically take place in any type of proliferating cell, they are only likely to have a clonal advantage in the context of a particular developmental pathway.8 This is obviously the case when pluripotent stem cells or committed progenitor cells are the target. Therefore it comes as no surprise that cancer cells not only share the self-renewal and differentiation capabilities of normal stem cells, but also utilize the same tissue-specific and tissue-dependent signalling pathways. Since stem cells are mostly non-cycling cells, they should not be particularly susceptible to DNA damage. However, their utilization during periods of increased cell demand such as during fetal development and the maturation and activation of the immune system, as well as during regenerative rebounds, might render them significantly more vulnerable. Moreover, the self-renewal machinery is already activated in stem cells and it may be much simpler and require fewer mutations to maintain it than to reactivate it completely.8
Stem cells are essential for embryogenesis and for the preservation of many tissues.30 The observation that many aspects of tumour development mimic organogenesis gone awry fuelled the idea that tumour growth and metastasis may be driven in a similar fashion by a small, but specialized, population of cancer stem cells. This hypothesis was further corroborated by the results of in vitro and in vivoclonogenic assays that were obtained with a variety of tumour tissues and cell lines. This fascinating and intellectually challenging concept has tantalizing and wide-ranging implications for cancer research as well as for cancer treatment.30 If confirmed, it will necessitate a shift in diagnostic and therapeutic endeavours from the currently more general quantitative to a very specialized qualitative level. It will probably prove more difficult to identify and characterize the properties of these comparatively rare cancer stem cells than it was to single out the normal ones. As in normal differentiating cell lineages, cancer stem cells can most probably be distinguished from most of their more harmless descendants only by virtue of subtle functional differences, and not on the basis of their shared rearranged genetic make-up. To cure cancer it would be sufficient to simply kill these cancer stem cells rather than to eliminate the whole bulk of neoplastic cells. A successful therapy would then appear as ‘spontaneous remission’. The concept of cancer stem cells also explains why even complete clinical and molecular remissions might spare cells that eventually cause a relapse. As is the case with normal stem cells in the physiologic situation, cancer stem cells could even prove to be more resistant to chemotherapeutics than their more ‘mature’ offspring. Despite their high sensitivity, the techniques currently employed for the detection of minimal residual disease would then prove insufficient to discriminate between these functionally heterogeneous cells. Finally, the shutting down or regression of transformed stem cells could also lead to spontaneous remissions, such as those regularly encountered in TMD and 4S-stage neuroblastoma.
The role of deterministic and stochastic information
Selective growth can arise in a cellularly autonomous fashion by mutation in the prospective cancer cell, by differences in the regulatory microenvironment, or more usually by some combination of both.2,31The environment provides deterministic and stochastic stimulating and inhibiting information to this process.26 Deterministic information comprises specific predetermined signals, for example growth factors and hormones, whereas stochastic information appears primarily as random noise. However, driven by chance, it can also be converted into signals under special circumstances. One example of such a process in a physiologic situation is the genetically determined generation of antibody diversity.26 Pre- and postnatal development, maturation, and challenge of the immune system, in particular, generate an enormous genetic diversity, and incidentally also provide the fateful repertoire of mutations that trigger the development of the various forms of childhood ALL. Screening of umbilical cord blood as well as the peripheral blood of healthy adults indicates that a variety of leukaemic fusion genes are generated much more frequently than the incidence of the respective leukaemias would suggest.32 This implies that one or more additional genetic hits and selective steps are required for the disease to develop. In the aetiology of childhood ALL, for example, the proliferative stress associated with deficient infectious exposure in infancy and an abnormal immune response to subsequent delayed common infections should constitute pivotal factors, according to Greaves' hypothesis.5 A well-established example of a stochastic trigger is the contribution of malaria to the development of endemic Epstein–Barr-virus-positive Burkitt lymphoma.
Childhood cancer: a disease of disturbed prenatal development
Solid tumours
As early as 1877, the German pathologist Julius Cohnheim first formulated the theory that most childhood malignancies arise when the delicate balance between growth, development, and differentiation of early fetal organogenesis is disrupted. This theory is based on the histopathologic demonstration of putative premalignant lesions at autopsy that resemble neuroblastoma and Wilms tumour (nephrogenic rests) and, more recently, on the detection of leukaemia-specific fusion genes in the umbilical cord blood and Guthrie cards of healthy neonates.4,5,8,32 The high frequency of such events (approximately 1–5 per cent) suggests that paediatric cancers may be initiated at a high rate, but with only a low risk of penetrance to malignancy.8,32 Depending on their particular genetic make-up, such premalignant lesions may regress, mature, or progress. Regression and maturation are features that are unique to certain childhood malignancies and virtually never encountered in adult neoplasms. One of the main reasons for this exclusivity may be that these faulty or delayed processes essentially mirror just two of the fundamental components of physiologic organ development.
Twin studies
The extraordinary opportunity that twins, and occasionally triplets, offer for the investigation of the developmental timing, natural history, and molecular genetics of paediatric leukaemia has recently been reviewed by Greaves.7 With the help of molecular markers, such as clone-specific IG and TCR gene rearrangements and unique fusion gene sequences, it has been shown that leukaemias in concordant twins have a common clonal origin. The explanation for this intriguing finding is that, at least in twins with a single monochorionic placenta, the pre-leukaemic cells spread from one twin to the other via a shared blood circulation. In other instances, this may occur via vascular anastomoses. However, postnatal latency can vary considerably between twins and the onset of leukaemia can occasionally be delayed by up to 14 years. This observation supports the notion that as yet undetermined factors must cause the final leukaemic transformation during the postnatal period.5
Leukaemias
Given the very young age at onset, a prenatal origin of infant leukaemia must of course be expected. Indeed, some cases are even diagnosed neoor prenatally. However, the surprising insight that we have now gained from an innovative series of ‘back-tracking’ experiments is that most childhood ALL and AML cases which are diagnosed in the first decade of life are also initiated in utero. 5,32 Even more surprising are the results of sophisticated IG and TCR gene rearrangement studies, which indicate that, except for leukaemias with an MLL/AF4 gene rearrangement, the first transforming hit must take place at a very early stage, i.e. between weeks 7 and 9 of gestation. The only genetically defined subtype for which a prenatal origin has not yet been proved is the more mature form of pre-B ALL with a t(1;19) and an E2A/PBX1 gene rearrangement.8
Implications for clinical management
The combined insights obtained from twin studies, Guthrie cards, and umbilical cord blood screening contribute significantly to our understanding of the biologic basis of leukaemias. These analyses have now unequivocally ascertained that childhood leukaemias in general originate from an early prenatal transforming event. Nevertheless, it is still unclear what exactly constitutes this first initiating step. To a large extent, the order in which the further sequences of events take place and how they ultimately lead to the development of these life-threatening diseases is also unknown. Furthermore, preliminary but compelling evidence suggests that concurrently generated presumptive preleukaemic cells coexist with overt leukaemic blasts at diagnosis and that they may survive chemotherapy and occasionally spawn later relapses.8 These discoveries have significant implications and consequences not only for the design, conduct, and interpretation of molecular genetic epidemiologic studies which examine causality, but also for PCR-based monitoring of MRD in childhood leukaemias.
Finally, the spectacular results of these investigations also provide the basis for risk assessment and counselling in cases of twins with asynchronous disease manifestations. This risk may approach 100% for identical monozygotic infant twins who shared a single monochorionic placenta. In parallel with the general decrease in the age-associated incidence rate, it then also declines to approximately 10 per cent in older children7 (see also Chapter 1). Because the calculated risks are still substantial, clinical surveillance, at least, of the healthy co-twin seems to be justified. However, a more thorough molecular scrutiny is definitely warranted if a co-twin is to be considered as a donor.
Appendix
Appendix. Websites with further information about haematologic neoplasms and solid tumours
|
References
1. Green DR, Evan GI (2002). A matter of life and death. Cancer Cell 1, 19–30.
2. Anonymous (2001). Nature insight: cancer. Nature 411, 336–95.
3. Breivik J (2001). Don't stop for repairs in a war zone: Darwinian evolution unites genes and environment in cancer development. Proc Natl Acad Sci USA 98, 5379–81.
4. Maris JM, Denny CT (2002). Focus on embryonal malignancies. Cancer Cell 2, 447–50.
5. Greaves M (1999). Molecular genetics, natural history and the demise of childhood leukaemia. Eur J Cancer 35, 1941–53.
6. Hanahan D, Weinberg RA (2000). The hallmarks of cancer. Cell 100, 57–70.
7. Greaves M, Maia AT, Wiemels JL, Ford AM (2003). Leukemia in twins: lessons in natural history. Blood, 102, 2321–33.
8. Greaves M, Wiemels J (2003). Origins of chromosome translocations in childhood leukaemia. Nat Rev Cancer, 3, 639–49.
9. Hood L, Galas D (2003). The digital code of DNA. Nature 421, 444–8.
10. Anonymous (2000). Nature insight: functional genomics. Nature 405, 820–65.
11. Anonymous (2003). The double helix—50 years. Nature 421, 396–453.
12. Jones PA, Baylin SB (2002). The fundamental role of epigenetic events in cancer. Nat Rev Genet 3, 415–28.
13. Steenman M, Westerveld A, Mannens M (2000). Genetics of Beckwith–Wiedemann syndrome-associated tumors: common genetic pathways. Genes Chromosom Cancer 28, 1–13.
14. Anderson GR, Stoler DL, Brenner BM (2001). Cancer: the evolved consequence of a destabilized genome. Bioessays 23, 1037–46.
15. Pihan G, Doxsey SJ (2003). Mutations and aneuploidy: co-conspirators in cancer? Cancer Cell 4, 89–94.
16. Balmain A (2001). Cancer genetics: from Boveri and Mendel to microarrays. Nat Rev Cancer 1, 77–82.
17. Maser RS, DePinho RA (2002). Connecting chromosomes, crisis, and cancer. Science 297, 565–9.
18. Perera FP (1997). Environment and cancer: who are susceptible? Science 278, 1068–73.
19. Huff V (1998). Wilms tumor genetics. Am J Med Genet 79, 260–7.
20. Garber JE, Offit K (2005). Hereditary cancer predisposition syndromes. J Clin Oncol 23, 276–92.
21. Albertson DG. Collins C, McCormick F, Gray JW (2003) Chromosome aberrations in solid tumors. Nat Genet 34, 369–76.
22. Rainis L, Bercovich D, Strehl S, et al. (2003). Mutations in exon 2 of GATA1 are early events in megakaryocytic malignancies associated with trisomy 21. Blood, in press.
23. Shannon K (1992). Genetic alterations in leukemia: events on a grand scale. Blood 80, 1–2.
24. Downing JR, Shannon KM (2002). Acute leukemia: a pediatric perspective. Cancer Cell 2, 437–45.
25. Brodeur GM (2003). Neuroblastoma: biologicalinsights into a clinical enigma. NatRev Cancer 3, 203–16.
26. Knudson AG (2001). Two genetic hits (more or less) to cancer. Nat Rev Cancer 1, 157–62.
27. Fodde R, Smits R (2002). Cancer biology. A matter of dosage. Science 298, 761–3.
28. Anonymous (2002). The chipping forecast. Nat Genet 32 (Suppl), 461–552.
29. Thilly WG (2003). Have environmental mutagens caused oncomutations in people? Nat Genet 34, 255–9.
30. Reya T, Morrison SJ, Clarke MF, Weissman IL (2001). Stem cells, cancer, and cancer stem cells. Nature 414, 105–11.
31. Rubin H (2001). Selected cell and selective microenvironment in neoplastic development. Cancer Res 61, 799–807.
32. Mori H, Colman SM, Xiao Z, et al. (2002). Chromosome translocations and covert leukemic clones are generated during normal fetal development. Proc Natl Acad Sci USA 99, 8242–7.
*… and, of course, also Mendel's Laws, which I take liberty to add as author of this chapter.
